Genetic regulation of RNA splicing in human pancreatic islets

Abstract

Genetic variants that influence transcriptional regulation in pancreatic islets play a major role in the susceptibility to type 2 diabetes (T2D). For many susceptibility loci, however, the mechanisms are unknown. We examined splicing QTLs (sQTLs) in islets from 399 donors and observed that genetic variation has a widespread influence on splicing of genes with important functions in islet biology. In parallel, we profiled expression QTLs, and used transcriptome-wide association and co-localization studies to assign islet sQTLs or eQTLs to T2D susceptibility signals that lacked candidate effector genes. We found novel T2D associations, including an sQTL that creates a nonsense isoform in ERO1B, a regulator of ER-stress and proinsulin biosynthesis. The expanded list of T2D risk effectors revealed overrepresented pathways, including regulators of G-protein-mediated cAMP production. This data exposes an underappreciated layer of genetic regulation in pancreatic islets, and nominates molecular mediators of T2D susceptibility.

Genetic regulation of RNA splicing in human pancreatic islets

“…This analysis also revealed a widespread impact of genetic variants on splicing variation, with 4858 cis-sQTLs at FDR ≤ 1%, 25% of which showed >10% shift in splice site usage in reference vs. alternate alleles (Fig. 1a, b, Additional file 3: Table S2 [21]). The 4858 sQTL junctions included alternative usage of 5′ exons, 3′ exons, mutually exclusive or skipped exons, or influenced combinations of S3); the bottom panel shows enriched annotations using EnrichR and Benjamini-Hochberg-adjusted p-values such splice variants (Fig.…”

“…Focusing on genes expressed in >10% of samples in each cohort, we found cis-eQTLs in 3433 genes (eGenes) at FDR ≤ 1% (Fig. 1a, Additional file 2: Table S1 [21]). This analysis also revealed a widespread impact of genetic variants on splicing variation, with 4858 cis-sQTLs at FDR ≤ 1%, 25% of which showed >10% shift in splice site usage in reference vs. alternate alleles (Fig.…”

“…A junction-based nominal p-value threshold was then calculated using pt and the beta distribution parameters from QTLtools. For each significant junction, variants with a nominal p-value below the junction-level threshold were considered in subsequent analyses (significant nominal cis-sQTL variants) [21].…”

Stem Cell-Derived β Cells: A Versatile Research Platform to Interrogate the Genetic Basis of β Cell Dysfunction

“…Many of the advances seen in β cell research during the last decade have relied on islets from donors. Unbiased -omic studies have explored the shape of the β cell genome identifying non-coding RNAs modulated in T2D [ 69 ], RNA splicing [ 70 ], regulatory elements [ 71 ], chromatin accessibility, DNA methylation [ 72 , 73 , 74 ] and three-dimensional chromatin architecture [ 75 ], findings that were integrated with GWAS signals. While large databases that integrate genomic information with tissue specific expression are a precious asset in identifying expression quantitative trait loci (eQTLs), these databases rarely contain information from pancreatic islets [ 76 ].…”

“…One of these is altered splicing, which can result from SNPs and impact disease risk of complex traits independently from effects on gene expression [ 314 ]. An atlas of splicing QTLs (sQTLs) across human islets was recently generated, and this constitutes an additional layer to explore the genetic basis of T2D risk [ 70 ]. The authors found sQTLs in previously reported T2D-risk loci but also found small-effect T2D-associations in new candidate genes, such as an exon-skipping variant of ERO1B .…”